Pencil

What Pencil Does Differently

Most AI ad tools generate creatives from a prompt and call it done. Pencil takes a different angle: it connects to your existing ad accounts, reads your historical performance data, and uses that to score and prioritize new creative concepts before you spend money testing them.

The pitch is straightforward—stop guessing which ad will work and let past winners inform the next batch. For teams running dozens of ad variations a month, that feedback loop has real value.

Where It Works Well

Pencil is genuinely useful for performance marketing teams at established DTC brands who already have a meaningful volume of ad data. If you've been running Facebook or TikTok ads for 12+ months and have clear winners and losers in your account, Pencil can surface patterns you might miss manually and generate new concepts that stay within proven creative territory.

The workflow is faster than briefing a designer for every iteration. You can produce a batch of variations quickly, see predicted performance scores, and prioritize what to test—cutting down the time between insight and execution.

Where It Falls Short

The ROI score of 24/100 reflects a real problem: the tool's value is gated behind data you may not have. Early-stage advertisers or brands with thin ad history get outputs that are essentially generic templates with your brand colors dropped in. The prediction engine has nothing to learn from.

Creative quality is also a consistent complaint. Outputs can feel formulaic—safe, on-trend formats rather than genuinely differentiated work. If your brand competes on creative distinctiveness, Pencil's outputs may need heavy editing before they're usable.

At $119/month minimum, the price-to-value ratio is hard to justify unless you're spending at least $5,000–$10,000/month on paid social. Below that threshold, the cost of the tool outweighs the iteration speed it provides.

Bottom Line

Pencil is a reasonable fit for mid-market performance teams who want to move faster on creative testing and already have the data to make the prediction layer meaningful. It's not a tool for startups, brand-led creative teams, or anyone who needs genuinely novel output. If you're in the right stage and running the right volume, it can shorten your testing cycles—but it won't replace strategic creative thinking.